C2Q Q6600 Unstable Past 3.1GHz

J

just_sum_dude

n00b
Joined
Nov 1, 2004
Messages
36
I am holding out for Ivy Bridge. I have been running my Q6600 @ 2.66GHz for the past couple of years, but recently decided to see how far I could push it. The thing is, I'm having a lot of trouble getting the damn thing stable past 3.10GHz. Loads into windows fine, but forces to reboot or bsods on the OCCT Linpack test. Temps are reasonable, max 62C which is a bit toasty admittedly but Linpack creates more heat than I'll ever see in real world applications. I've increased vcore as high as 1.3825, with no real improvement. Motherboard is a Gigabyte X48-DS5. Ram is Mushkin Ascent XP2-8500 5-5-5-15 running at >1000Mhz. I realize that asking "why it no go fast" is kinda ridiculous considering the number of variables that can lead to instability. But I would appreciate any advice you guys could provide. Thanks.

- Nim
 
Try pushing the cpu voltage, 1.45 is max safe limit from Intel. As long as you keep the temps below 80 degrees, should be fine. Increase your VTT by .1v and northbridge voltage by .1v as well see if that helps.

Also try downclocking the memory just in case it is getting in the way of your OCing.
 
I needed to adjust the voltages of a lot more settings in the bios than just the vCore to get my Q6600 clocks stable (3.6GHz).

Just vCore may not be enough. I don't know anything about your board though and your bios options (unless you list them) so I can't tell you what you may need.
 
this is the go-to guide for q6600 overclocking...
http://hardforum.com/showthread.php?t=1198647&highlight=q6600+overclock+guide

it's about as comprehensive as it gets.

just my 2 cents based on the info you've given, might try upping the voltage on the North Bridge a notch or two to stabilise higher fsb's (344x9). vdroop might also be an issue so you may want to look at enabling Load Line Calibration (or whatever gigabyte calls it).

also take the RAM out of the equation by dropping the multiplier.

if your gonna do it properly you'll need to do only one step at a time. for example drop the CPU multi to 6 and push the FSB till it becomes unstable then increase NB volts. you keep doing that till you reach either your desired FSB for your final clock (355x9=3200 / 377x9=3400 / 400x9=3600 etc) or you reach the limit of the NB chip. shouldn't need to ask that you check the temps on CPU/NB while doing this.

what stepping your CPU is will affect you OC (B3/G0).

in any case, check the link, it covers everything.
 
Thanks guys. I've got it stable (6+ hours in OCCT) at 3.2GHz right now, spent a couple hours last night trying to get 3.3GHz stable, no such luck however. I think I'd be happy with 3.3GHz or 3.4GHz, but it's going to require some more trial and error.

It's G0 stepping, just for the record.

My settings for 3.2GHz are as follows:

Robust Graphics Booster: Auto
CPU Clock Ratio: 9x
CPU Host Clock Control: Enabled
CPU Host Frequency: 356
PCI Express Frequency: 100
C.I.A.2: Disabled
Performance Enhance: Turbo
System Memory Multiplier: 2.40B
Memory Frequency: 855

DRAM Timing Selectable - Manual
CAS Latency Time - 5
DRAM RAS# to CAS# Delay - 5
DRAM RAS# Precharge - 5
Precharge Delay - 18
Other Ram Settings - Auto

CPU/PCIEX Clock Driving Control - 800mV
CPU Clock Skew Control - Normal
(G)MCH Clock Skew Control - Normal

System Voltage Control - Manual
DDR2 OverVoltage Control - +0.3
PCI-E OverVoltage Control - +0.1
FSB OverVoltage Control - +0.1
(G)MCH OverVoltage Control - +0.1
Loadline Calibration - Enabled
CPU Voltage Control - 1.38725V
Normal CPU Vcore - 1.3250V
Click to expand...

Any advice on what I could do to push it farther?
 
The chip need more volts, you can go up to 1.45v although temps will be high.

Test in increments of .025v and see how much you need for stability and then decide if it is worth it.
 
You need to downgrade the memory frequency to 533MHz before overclocking procedures.
Then You can reach 4GHz by raising CPU Frequency to 445MHz with vcore 1.5v
Ofcourse with appropriate cooling.
 
New settings, still 3.20GHz.

Robust Graphics Booster: Auto
CPU Clock Ratio: 8x
CPU Host Clock Control: Enabled
CPU Host Frequency: 401
PCI Express Frequency: 100
C.I.A.2: Disabled
Performance Enhance: Standard
System Memory Multiplier: 2.00D
Memory Frequency: 802

DRAM Timing Selectable - Manual
CAS Latency Time - 5
DRAM RAS# to CAS# Delay - 5
DRAM RAS# Precharge - 5
Precharge Delay - 15
Other Ram Settings - Auto

CPU/PCIEX Clock Driving Control - 800mV
CPU Clock Skew Control - Normal
(G)MCH Clock Skew Control - Normal

System Voltage Control - Manual
DDR2 OverVoltage Control - +0.3
PCI-E OverVoltage Control - Normal
FSB OverVoltage Control - +0.15
(G)MCH OverVoltage Control - +0.15
Loadline Calibration - Enabled
CPU Voltage Control - 1.38725V
Normal CPU Vcore - 1.3250V
Click to expand...

Might increase the Vcore and pop the multi up to 9x, haven't decided yet.

- Nim
 
What about making
System Memory Multiplier:2B ( to become 533MHz before setting the CPU Host Frequency to 400)
and Loadline calibration: Disabled
and return the CPU Clock ratio to 9x of-course
 
Why should I want to use the B-Strap as opposed to the D-Strap?

From what I've read about Loadline Calibration, there's no real consensus on whether or not it's a "good thing".

I have a feeling that 3.60GHz is going to be unattainable for me. I'm not comfortable pushing the VCore very far past 1.40V. It's primarily a adequate cooling issue. North of 1.40V my temperatures really start to spike in OCCT Linpack (admittedly hotter than any real-world applications), too close to 70C for comfort.

My HSF.

Ran some synthetics, scores seem pretty good to me:

WPrime - 32M 14.711 sec

Geekbench 2.20 - Score 5891

Hyper Pi - 20.404 sec

3DMark Vantage - CPU Score 10934

Any further advice would be appreciated obviously.

- Nim
 
Last edited:
no matter what strap you use .. I assume B for 266 and it could be A in your motherboard that equal to 266MHz

and for your question "Why?' that to keep the ratio 1:1 DRam:FSB as there is no benefit to run the memory faster than the FSB in the Core2 Architecture (many of us know that)

Loadline calibration will keep the stability to the CPU voltage and that what we do not want .. We want the whole electricity flow to the CPU and when you reach your overclock target you could turn it Enable again..

and for your feeling about unattainable .. you will not lose a thing to just try reach your goal 3.6GHz as your CPU could stand for 1.5v and return to whatever you want.

Best Regards.
 
B is the 333Mhz Strap, D is the 400Mhz Strap, and I don't really care what the other two are. 266Mhz and 233Mhz I'd imagine. Regardless, I am already running it 1:1.
 
My experience is that the system was much more unstable with the 400MHz strap. I got much higher overclocks on straps A-C.
 
Just an update:

Currently stable @ 3.33GHz, pretty well satisfied, don't expect I'll push it any farther.

Settings:

Robust Graphics Booster: Auto
CPU Clock Ratio: 9x
CPU Host Clock Control: Enabled
CPU Host Frequency: 370
PCI Express Frequency: 100
C.I.A.2: Disabled
Performance Enhance: Standard
System Memory Multiplier: 2.40
Memory Frequency: 888

DRAM Timing Selectable - Manual
CAS Latency Time - 5
DRAM RAS# to CAS# Delay - 5
DRAM RAS# Precharge - 5
Precharge Delay - 15
Other Ram Settings - Auto

CPU/PCIEX Clock Driving Control - 800mV
CPU Clock Skew Control - Normal
(G)MCH Clock Skew Control - Normal

System Voltage Control - Manual
DDR2 OverVoltage Control - +0.3
PCI-E OverVoltage Control - +0.10
FSB OverVoltage Control - +0.15
(G)MCH OverVoltage Control - +0.225
Loadline Calibration - Enabled
CPU Voltage Control - 1.4250V
Normal CPU Vcore - 1.3250V
Click to expand...

Thanks for the advice/links/etc.

- Nim
 
You must log in or register to reply here.
Back
Top